May 5, 2026 | The diagnosis and management of hypertension is in a troublingly bleak state, despite myriad treatments and regular blood pressure checks in primary care settings intended to catch the condition early. The conundrum is rooted in a trio of problems that co-exist like “gears with broken teeth that don’t mesh together very well,” according to Michael Liebman, Ph.D., managing director of IPQ Analytics.
In medicine today, patients naturally expect an accurate diagnosis and effective treatment, but physicians may be constrained by standards of care and clinical guidelines as well as reimbursement policies, while the disease itself remains poorly understood, he points out. One of the core issues is that clinical researchers don’t understand how to stratify hypertension and thus doctors don’t fully appreciate or act on its complexity.
That’s why IPQ Analytics is reevaluating the diagnostic criteria for hypertension, relying more on real-world data than the more homogenized variety produced by clinical trials, Liebman says. The company’s process involves untangling the natural circadian pattern of blood pressure to personalize the condition as well as separating cardiac from vascular causes to ensure patients get the right treatment.
Liebman will be making a presentation about the novel diagnostic approach at the upcoming Next Generation Dx Summit in Washington, D.C. The backdrop is the grim reality that in the U.S. alone hypertension diagnosis and treatment costs more than $30 billion annually, involves nearly half the population, and is only controlled in about 25% of patients.
In the U.S., hypertensive disorders of pregnancy—chronic hypertension, gestational hypertension, preeclampsia, and eclampsia—occur in 13% to 16% of all pregnancies. They’re associated with 7% of all maternal deaths, and this can range to 30% in Black populations.
For IPQ Analytics, the starting point is preeclampsia. The aim is to identify hypertension risk before a woman becomes pregnant so it can be better managed and learn which types are more likely to progress to the pregnancy disorder, Liebman says.
Being an underserved problem, he adds, the hope is that the Food and Drug Administration (FDA) will eventually consider the use of novel alternatives to the blood pressure cuff, which is the agency’s current required reference standard for clearance of new non-invasive diagnostic devices for hypertension. IPQ Analytics is initially partnering with Cosinuss GmbH, based in Munich, Germany, which is developing a sensor that estimates blood pressure by analyzing pulse wave characteristics from blood flow in the ear
Blood pressure cuffs have a few key downsides, says Liebman, one of which is that American Medical Association guidelines for their use are widely ignored by physicians. Blood pressure readings can also vary by 10 to 15 points (mmHg) throughout the day as the body responds to activity, stress, and its internal clock.
“There’s a great deal of inaccuracy in what’s being measured, along with the fact that the guidelines are constantly changing internationally,” Liebman says. One reason for the guideline updates is that blood pressure cuffs are measuring “an anecdotal observation at a single point in time” while blood pressure is known to fluctuate by as much as 20% over the course of 24 hours.
Personalizing a Diagnosis
Having a mixed pharma and academic background, Liebman understands how clinical trial patients are recruited, the pressures in setting up studies, and why participants don’t accurately reflect real-world patients. Even with this selection bias, 90% of drug candidates that enter clinical trials ultimately fail to receive regulatory approval, many because of failure to adequately stratify the disease itself.
Even more concerning to Liebman, because he does systems modeling, is that the drugs getting approved invariably only work in 30% to 40% of the patients taking them, he says. “That says to me you’re not even targeting the right patients, or disease subtype, to start with.” The advances in molecular characterization and technology have been driving “precision medicine,” but without better understanding of the disease process, he questions whether they have also been driving “accurate medicine.”
To remedy the situation, the disease modeling process of IPQ Analytics “starts in the clinic and works backwards, not relying solely on the molecular measurements we make and then driving it forward,” he says. Part of the challenge when it comes to hypertension is the use of “umbrella” (primary and secondary) diagnoses rather than personalizing the disease.
The company consequently doesn’t use claims data. Although that accounts for much of the available real-world data, it’s “generated under pressure for business reasons,” says Liebman. “It’s not there to describe the disease at a more fundamental level.”
Disease is a process and not a state, he emphasizes, meaning clinicians don’t know where along the disease path patients are when they come in for a diagnosis. “They make a guess based on their experience and what little information has been published and what tests they can run.” The difficulty is that “two patients may present identically with the same lab results and the same symptoms but have completely different diseases ... [based on] when they appear for diagnosis in the course of their disease.”
What physicians could use are metrics that define the course of a disease over time (trajectory) to better understand where a patient currently sits on that progression vector (staging), and how rapidly the condition is progressing (velocity), says Liebman. In addition to closing those gaps for physicians, IPQ Analytics also looks to bridge the divide between real-world populations and participants in trials.
For one failed clinical trial with 6,000 participants over six years, Liebman offers as an example, no conventional method of subgroup analysis turned up any results until IPQ Analytics stepped in to stratify the condition and show the study sponsor it had unknowingly enrolled five different disease-expressing patient populations (from the 2022 paper showcased on the Institutional Research Information System). The value of the drug would otherwise have been “lost in that heterogeneity” as well as failing to identify that segment of patients who would truly benefit from its development.
The real-world data for this sort of work is acquired on a disease-by-disease basis based on the modeling task. In the case of perimenopause, where the focus is on progression, data is needed starting during puberty, says Liebman, and the company has initiated international collaborations to collect this data.
For its work in breast cancer, it is not enough to work with some of the advanced breast cancer centers that only see patients for their post-diagnosis treatment. IPQ Analytics is interested in looking for disease insights based on patients’ lives before they are diagnosed with the disease in addition to their response to treatment.
Even the most prominent oncologists and breast surgeons commonly conflate the ideas of early detection and prevention, he notes. “They’re not getting to the cause ... actually referring to the prevention of disease progression, which is very important but it’s not addressing disease prevention.”
That is, women are living longer with breast cancer, but the incidence is going up. “We’re not preventing it; we’re treating it better, and we’re diagnosing it earlier,” he says. “We could have a much bigger impact if we can start to trace back to the root causes.”
For the purposes of studying preeclampsia, the partnership with Cosinuss aims to accurately capture the circadian pattern of blood pressure and thereby get a better reading than is possible with existing technologies. Today, if a pregnant woman exhibits hypertension as measured by a standard blood press cuff in the doctor’s office, she’ll typically be given a portable blood pressure monitoring cuff to wear for perhaps 24 to 48 hours. The battery-powered device automatically inflates and takes a reading every 30 minutes.
Although IPQ Analytics has access to about 30,000 of those records, the data doesn’t necessarily recognize that being awakened from sleep when the cuff starts inflating can disrupt a person’s blood pressure, as can anticipating the cuff inflating during the day. “It’s not really measuring an accurate blood pressure in a naïve situation,” he says.
The Cosinuss device is an earpiece that uses an optical measuring method based on signals tied to the blood volume in the vessels in the ear canal. “It doesn’t give you the same measurement value that you get with the cuff, but it is measuring another property of the blood flow that is related to the elasticity of your blood vessels, which is the intent of blood pressure monitoring,” explains Liebman.
Hypertension can be of either cardiac or vascular origin and has about 10 different contributing processes for which different classes of anti-hypertensive medicines have been developed, he continues. The collaborative study will be looking to manage only a couple of them with the current arsenal.
What physicians tend to do is basically go down the list of anti-hypertensives and “experiment” on patients to see what they respond to, says Liebman. “When none of them work, they move onto drug combinations.” Experiments in simple chrono-dosing have also been attempted without evaluation of the detailed circadian fluctuations that an individual patient may exhibit.
Liebman says that continuous blood pressure monitoring with the earpiece should reveal different waveform patterns and how they’re related to the underlying pathway, to enable patients’ blood pressure to be personalized so doctors know the best drug to prescribe. “If no drugs are currently directed toward that particular process, that gives us new targets for drug development.”
The study population will include pregnant women in Mexico who have and haven’t had preeclampsia or been diagnosed as hypertensive, he says. Mexico’s federal health regulatory agency is interested in solving the clinical problem and expected to more readily embrace the use of cuffless devices in clinical trials than the FDA.
Ultimately, multiple partners will be needed to come up with a means to measure disease stage and the pace of progression to guide treatment choice and monitor patient response. The key challenge is that diagnostic measurements are being made using technology developed without an understanding of the complexity of hypertension, says Liebman. “Without that understanding, we’re more attuned to using the technology we have to make a measurement and then trying to figure out how that measurement relates to a complex disease process.”
The objective of IPQ Analytics is to identify “relatively simple, noninvasive” observations early on that can help guide the clinical management of patients with diseases based on how they evolve over their lifetime, he says. Achieving that ambitious goal will require patients willing to participate in studies involving wearable devices, including those resembling an earpiece they’d wear at the gym, to collect the necessary real-world data.
One big opportunity with the forthcoming preeclampsia study is that it will prove the value of cuffless blood pressure devices, so that the FDA might decide to drop its expectation that they reproduce the type of measurements provided by the traditional oscillometric method, says Liebman. While they both display the standard two-number readout, they’re not measuring identical physiological markers.
Liebman is now launching a nonprofit research institute, known as Woven, which will be “studying how women go through perimenopause uniquely ... what to expect, what’s normal and not, and how to manage it better” as well as how the transition “starts as early as puberty in terms of putting a woman on a track ... personalizing a woman’s risk for breast cancer or cardiovascular disease or osteoporosis,” he reports. Rather than tying the discovery process to pre-existing data and accommodating statistical tools, it will pursue the creation of model-driven data in studies to be conducted in Australia, Mexico, and Chile in a quest to meet medicine’s need for causality rather than correlation.
Woven will separately be studying hypertensive disorders of pregnancy with a systems approach looking at conditions that present at different points across the lifespan of women and have unknown or poorly understood and separable long-term effects, he says. Among the kinds of data to be collected are the hormones produced by the fetus and the mother over the course of pregnancy that are maximized at 40 weeks.
One outstanding question is the impact of early deliveries, now viable in as few as 22 weeks, on the perimenopause transition and breast cancer risk later in life. “The challenge is their medical records don’t normally show the gestational age of the delivery, so that’s the kind of data we’re collecting ... internationally.”
Many standard clinical practices reflect current standards of care with physicians not positioned to question their adequacy or be involved in clinical research, notes Liebman. “We should be challenging them constantly” to reveal the underlying complexities—of the patient, the disease, and how medicine is practiced—to make complete solutions achievable.